Environment mapping using probabilistic quadtree for the guidance and control of autonomous mobile robots

Abstract

This paper presents the implementation of an environment mapping technique based on a probabilistic quadtree which records the location of static and dynamic obstacles, as well as the certainty over each obstacle's estimated position. The quadtree-based map is updated online (i.e. near-real time) based on multi-sensor feeds originating from one to three X80 mobile robots operating simultaneously. The probabilistic quadtree map is part of a guidance and navigation control system which combines a Genetic Algorithm-based global path planner and a Potential Field local controller. The centralized map is shared by all mobile robots although each robot has an independent controller. Performance of the proposed method for this guidance and navigation control system is demonstrated experimentally with the Dr. Robot's ™ wireless X80 mobile robots.